Abstract: Aiming at the rapid burning and wide range of forest and grassland surface fires, and the difficult problem of navigation and fire source localization of fire-fighting robots in complex environments, a forest and grassland pneumatic fire-fighting robotic system based on robot operating system（ROS） is designed. The hardware part of the system consists of a tracked robot chassis, fan system, wind turbine control mechanism, and a combination of inertial guidance, camera and LIDAR and other sensors, using GPS satellite positioning information and AMCL algorithm fusion to achieve the robot's localization. Proposing a fuzzy logic-based dynamic weighted A* algorithm to generate the environmental complexity index through fuzzy reasoning on the density of the obstacles and the distribution of the discrete degree in the 25% obstacle density of 30×30 raster map simulation experiments, compared with the traditional A* algorithm, the number of search nodes decreased by about 82.1%, the planning time was shortened by about 52.8%, and the total generation value was reduced by about 1.8%; in the 45% obstacle density of 30×30 raster map simulation experiments, compared with the traditional A* algorithm, the number of search nodes was reduced by about 24.4%, and the total generation value was reduced by about 3.4%, the The planning time is reduced by about 22.3%. The dynamic obstacle influence factor is introduced into the DWA evaluation function, which can accurately reach the target point in a shorter time compared with other algorithms under the premise of guaranteeing the safety; and the blower angle control strategy with flame height feedback is established, which can accurately locate the fire source and adjust the angle of the blower to extinguish the fire. The results show that the system is able to achieve autonomous obstacle avoidance and target positioning in the complex woodland environment, and can meet the demand for fire extinguishing in the complex environment of forest and grassland.

Key words: ROS, pneumatic fire-fighting robot, A* algorithm, DWA algorithm, path planning